Methodology description of the labour force sample survey                  

SWITZERLAND 1997

Enquête suisse sur la population active (ESPA)/

Schweizerische Arbeitskräfteerhebung (SAKE), 1997

Responsibility for methodology, organization and co-ordination:

BFS OFS UST

Bundesamt für Statistik - Office fédéral de la statistique

Ufficio federale di statistica - Uffizi federal da statistica

 

Implementation of the interviews in the laboratories of LINK (Luzern) and MIS (Lausanne)

a) geographical: persons permanently resident in Switzerland (during the whole year)
b) persons covered: persons aged 15 or more who are Swiss citizens or foreign citizens with a residence permit. Excluded are the individuals living in Switzerland during a short period (for example, during the summer), the cross-border workers and the refugees.

Given the chosen survey method (through phone interviews), only the household that dispose of a private telephone are taken into consideration.

The collective households (hospitals, prisons, residential centers, boarding schools, old people’s homes, nursing homes, hostels, etc.) are considered only if the individuals living there dispose of their own private telephone.

The Survey is carried out once a year, during the 2nd quarter (April-June).

The week before the survey.

The Swiss Labour Force Survey, existing since 1991, provides important, internationally comparable information on employment in Switzerland. Every year approximately 16,000 persons chosen at random from the telephone register of the Swiss PTT are interviewed over the phone. Participation is voluntary. The results of the sample survey are projected onto the whole population and published.
In the 15-minute interviews, questions are asked on profession, professional experience, working times, working conditions, professional training, job seeking, former occupation, reasons for not being economically active and income. Likewise on the division of labour among branches and professional groups. The family environment and topics such as the rent of flats and care of children are taken into account. Data privacy is guaranteed.
The main goal of the LFS are:

• to provide every year representative data on the socio-economic structure  and the participation in the labour market of the Swiss population;
• to give a good picture of the employed, unemployed and non-active individuals, based on economic and socio-demographic criteria;
• to monitor the structural change of the labour market;
• to define - with the help of the panel - the main factors influencing the participation to the labour market and their consequences on the social conditions of the individuals.

Many variables that the EU countries are required to include in their LFSs are included in the Swiss LFS as well. Suggestions by the ILO have been taken into consideration when the questionnaire was designed. Moreover, a list of country-specific variables have been included, about for example the military service, the voluntary work, the education and training of the employees, and the child care.

The Swiss Labour Force Survey applies the ILO definitions:

 

Employed: All persons aged 15 or over who have worked at least one hour in the reference week (wage-earners or self-employed) or who have worked in a family business without pay. All apprentices are also included. Moreover, employed are also those who have not worked in the reference week, but who are officially hired by an employer (and were absent because of illness, holidays, etc.). Included in the definition are therefore:

• all forms of atypical work (working at home, working in private houses - child caring, cleaning, private lessons, etc.)
• all forms of occasional work

Unemployed: The unemployed (in accordance with international norms) are persons aged 15 or over who are not gainfully employed at the time of the survey, who are actively looking for a job (in the four weeks preceding the Survey) and are basically available, thus not seriously ill, for example (who can start working in the four weeks after the Survey). On the basis of differing terminology, the number of unemployed in the LFS tends to be higher than the number of unemployed registered at the employment exchange recorded by the Federal Office for Industry and Labour.

 

Not economically active: Not economically active are all the remaining persons aged 15 and over, viz. those who are exclusively involved in education and further education or training (thus do not have a casual job), housewives and men staying at home and taking care of the household as well as old age pensioners. Every form of domestic work and voluntary work are therefore considered as "non-active".

 

For practical purposes the sample is divided in 8 sub-groups:

Employed:
1) wage-earners;
2) self-employed;
3) apprentices;
4) soldiers
Unemployed:
5) unemployed without a sure job;
6) unemployed with a sure job;
Inactive:
7) inactive with a previous paid activity;
8) inactive without a previous paid activity.

Every one of these sub-groups is asked a different set of questions.

The reference person
This is a very important concept, because the Swiss LFS differs from all other countries' surveys in this respect.
The reference person is randomly chosen among all the household members who are at least 15. The reference person - who is always the same in the occasion of all subsequent interviews - is required to answer all the questions in the questionnaire. That is, detailed information is collected only for the reference person. Only some limited information is collected for the other household members by asking the reference person. The reference person cannot be replaced by another household member, except when he/she is absent during the 3 months that the survey lasts (for example, he/she is abroad). If the reference person cannot answer because he/she is ill or very old, an indirect interview is taking place (so called "interview proxi"), i.e. another household member answer for him/her (giving answers that refer to the reference person).

The occupation classification

The industry classification

The initial sample is composed by 20,000 to 22,000 addresses. The actual interviews are instead 16,000 to 18,000. The drop-outs depend essentially on two reasons:

1. the participation to the survey is not compulsory, but facultative;
2. some randomly chosen phone numbers are not valid, some households may not be reached, etc.

The sample cannot be exploited for cantonal studies given the reduced sample sizes at this level. For this reason the CSO has decided to exceptionally increase the sample size in 1995 (the year in between two censuses of 1990 and 2000).
The addresses are taken from the telephone book. A number of addresses proportional to the number of inhabitants and the response rate is randomly selected for each canton. The cantons that contribute to the financing of the LFS (Zürich, Vaud, Genève) have a bigger sample size.
According to ad-hoc studies, the sample is not significantly biased by the fact that 5% of the population is excluded from the population of origin because of lack of private phone.

 

Panel structure: The Swiss LFS is a rotating panel. Every year one fifth of the households already included in the sample is replaced; the other four fifths are re-interviewed. Every household, therefore, stays in the sample for 5 consecutive years.

The Survey is carried out by phone interview aided by the computer.

 

The response rate is 70% for the first interviews and 87% for the panel interviews.
The households that couldn't be contacted was in 1995 3.9% (first interview) and 1.9% (panel interview). There is no limit to the number of attempts to be made for the interview.
The refusal rate is about 16% (first interview) and 3% (panel interview).
The neutral losses (invalid phone numbers, collective households, second houses, company phone numbers) represent 2-4% (first interview) and 1-2% (panel interview).
The other losses (households that was impossible to contact, died persons, linguistic problems, etc.) represent about 5-6% in both types of interview.

 

 

The sample weights are computed in three steps:
1) correction for the different probabilities of being included in the sample
A weight equal to the inverse of the probability of being included in the sample is assigned to each individual. The criteria taken into account include the region of residence, the number of telephone lines in the households, the number of household members.
2) only for the panel: correction for the probability of drop-out.
This probability is estimated using a logit model, including as explanatory variables the personal characteristics of the reference person (age, sex, etc.) and the number of previous inclusions in the LFS. The estimated probability is then combined with the inclusion probability (see previous point).
3) post-stratification correction: correction needed to make the sample correspond to the total Swiss population (correction for losses and non-responses at the first interview and for the fact that not the whole Swiss population has a telephone line).
The correction is made by a procedure of post-stratification in three subsequent steps, according to the region (first steps), to the marital status (second step) and to the age, sex and origin - Swiss or foreign - (third step). In order to get a solution that satisfies the three criteria simultaneously, the procedure is iterated several times. The post-stratification is based on a permanently resident population aged at least 15. The extrapolation is then made on this population. In other words, each person included in the LFS corresponds to 320 "real" persons.

The first results of the Swiss Labour Force Survey are presented during October-November as press releases in German, French and Italian.

The detailed results, including tables and comments, are published at the beginning of the year by the Federal Statistical Office. This is a bilingual publication (in German and French) in the series "Erwerbsleben/Emploi et vie active": Die Schweizerische Arbeitskräfteerhebung (SAKE)/L' enquête suisse sur la population active (ESPA), Kommentierte Ergebnisse und Tabellen/Résultats commentés et tableaux, Bundesamt für Statistik/Office fédéral de la statistique

 

Every year a leaflet - whose title is "Vademecum" - is also published. It includes simple and straightforward information about the main goals of the LFS as well as some basic results. This leaflet is distributed to all sampled households.

 

In the "Statistischen Jahrbuch der Schweiz/Annuaire statistique de la Suisse" and the "Taschenstatistik der Schweiz/Mémento statistique de la Suisse" some data obtained from the LFS are included as well.

 

The LFS data make the object of specific studies and detailed analyses which are published as papers apart.